Reliability and availability in reconfigurable computing: a basis for a common solution

Dynamically reconfigurable SRAM-based field-programmable gate arrays (FPGAs) enable the implementation of reconfigurable computing systems where several applications may be run simultaneously, sharing the available resources according to their own immediate functional requirements. To exclude malfunctioning due to faulty elements, the reliability of all FPGA resources must be guaranteed. Since resource allocation takes place asynchronously, an online structural test scheme is the only way of ensuring reliable system operation. On the other hand, this test scheme should not disturb the operation of the circuit, otherwise availability would be compromised. System performance is also influenced by the efficiency of the management strategies that must be able to dynamically allocate enough resources when requested by each application. As those resources are allocated and later released, many small free resource blocks are created, which are left unused due to performance and routing restrictions. To avoid wasting logic resources, the FPGA logic space must be defragmented regularly. This paper presents a non-intrusive active replication procedure that supports the proposed test methodology and the implementation of defragmentation strategies, assuring both the availability of resources and their perfect working condition, without disturbing system operation.

Floating-point exponentiation units for reconfigurable computing

The high performance and capacity of current FPGAs makes them suitable as acceleration co-processors. This article studies the implementation, for such accelerators, of the floating-point power function xy as defined by the C99 and IEEE 754-2008 standards, generalized here to arbitrary exponent and mantissa sizes. Last-bit accuracy at the smallest possible cost is obtained thanks to a careful study of the various subcomponents: a floating-point logarithm, a modified floating-point exponential, and a truncated floating-point multiplier. A parameterized architecture generator in the open-source FloPoCo project is presented in details and evaluated.

Introduction to Special issue on Reconfigurable computing and FPGAs

During the last three decades, FPGA technology has quickly evolved to become a major subject of research in computer and electrical engineering as it has been identified as a powerful alternative for creating highly efficient computing systems. FPGA devices offer substantial performance improvements when compared against traditional processing architectures via custom design and reconfiguration capabilities.

Reconfigurable computing for Monte Carlo simulations: results and prospects of the Janus project

We describe Janus, a massively parallel FPGA-based computer optimized for the simulation of spin glasses, theoretical models for the behavior of glassy materials. FPGAs (as compared to GPUs or many-core processors) provide a complementary approach to massively parallel computing. In particular, our model problem is formulated in terms of binary variables, and floating-point operations can be (almost) completely avoided. The FPGA architecture allows us to run many independent threads with almost no latencies in memory access, thus updating up to 1024 spins per cycle. We describe Janus in detail and we summarize the physics results obtained in four years of operation of this machine; we discuss two types of physics applications: long simulations on very large systems (which try to mimic and provide understanding about the experimental non equilibrium dynamics), and low-temperature equilibrium simulations using an artificial parallel tempering dynamics. The time scale of our non-equilibrium simulations spans eleven orders of magnitude (from picoseconds to a tenth of a second). On the other hand, our equilibrium simulations are unprecedented both because of the low temperatures reached and for the large systems that we have brought to equilibrium. A finite-time scaling ansatz emerges from the detailed comparison of the two sets of simulations. Janus has made it possible to perform spin glass simulations that would take several decades on more conventional architectures. The paper ends with an assessment of the potential of possible future versions of the Janus architecture, based on state-of-the-art technology.

Reconfigurable Architectures for General-Purpose Computing

General-purpose computing devices allow us to (1) customize computation after fabrication and (2) conserve area by reusing expensive active circuitry for different functions in time. We define RP-space, a restricted domain of the general-purpose architectural space focussed on reconfigurable computing architectures. Two dominant features differentiate reconfigurable from special-purpose architectures and account for most of the area overhead associated with RP devices: (1) instructions which tell the device how to behave, and (2) flexible interconnect which supports task dependent dataflow between operations. We can characterize RP-space by the allocation and structure of these resources and compare the efficiencies of architectural points across broad application characteristics. Conventional FPGAs fall at one extreme end of this space and their efficiency ranges over two orders of magnitude across the space of application characteristics. Understanding RP-space and its consequences allows us to pick the best architecture for a task and to search for more robust design points in the space. Our DPGA, a fine- grained computing device which adds small, on-chip instruction memories to FPGAs is one such design point. For typical logic applications and finite- state machines, a DPGA can implement tasks in one-third the area of a traditional FPGA. TSFPGA, a variant of the DPGA which focuses on heavily time-switched interconnect, achieves circuit densities close to the DPGA, while reducing typical physical mapping times from hours to seconds. Rigid, fabrication-time organization of instruction resources significantly narrows the range of efficiency for conventional architectures. To avoid this performance brittleness, we developed MATRIX, the first architecture to defer the binding of instruction resources until run-time, allowing the application to organize resources according to its needs. Our focus MATRIX design point is based on an array of 8-bit ALU and register-file building blocks interconnected via a byte-wide network. With today's silicon, a single chip MATRIX array can deliver over 10 Gop/s (8-bit ops). On sample image processing tasks, we show that MATRIX yields 10-20x the computational density of conventional processors. Understanding the cost structure of RP-space helps us identify these intermediate architectural points and may provide useful insight more broadly in guiding our continual search for robust and efficient general-purpose computing structures.

Compiling Techniques for Coarse Grained Runtime Reconfigurable Architectures

In this paper we develop compilation techniques for the realization of applications described in a High Level Language (HLL) onto a Runtime Reconfigurable Architecture. The compiler determines Hyper Operations (HyperOps) that are subgraphs of a data flow graph (of an application) and comprise elementary operations that have strong producer-consumer relationship. These HyperOps are hosted on computation structures that are provisioned on demand at runtime. We also report compiler optimizations that collectively reduce the overheads of data-driven computations in runtime reconfigurable architectures. On an average, HyperOps offer a 44% reduction in total execution time and a 18% reduction in management overheads as compared to using basic blocks as coarse grained operations. We show that HyperOps formed using our compiler are suitable to support data flow software pipelining.

Integer Sequence window based Reconfigurable FIR filters

Conventional hardware implementation techniques for FIR filters require the computation of filter coefficients in software and have them stored in memory. This approach is static in the sense that any further fine tuning of the filter requires computation of new coefficients in software. In this paper, we propose an alternate technique for implementing FIR filters in hardware. We store a considerably large number of impulse response coefficients of the ideal filter (having box type frequency response) in memory. We then do the windowing process, on these coefficients, in hardware using integer sequences as window functions. The integer sequences are also generated in hardware. This approach offers the flexibility in fine tuning the filter, like varying the transition bandwidth around a particular cutoff frequency.

Metacomputing on clusters augmented with reconfigurable hardware

Accepted Version